The Inbox: Morphic-cmfcmf.1490.mcz

Am 03.07.2019 02:46:45 schrieb David T. Lewis <[hidden email]>:

tl;dr Is there any reason NOT to move/merge Morphic-pre.1491 and
Morphic-cmfcmf.1490 from inbox to trunk?

Descending from Morphic-pre.1489, we have Morphic-pre.1490 that was moved
from inbox to treated, and we also have Morphic-cmfcmf.1490 still in inbox.
Parallel to that, we also have inbox Morphic-pre.1491, with inbox Morphic-ct.1492
descended from it, and Morphic-ct.1492 depending on inbox Tools-ct.851 (and
presumably also Compiler-ct.407 and Tests-ct.412.

I have not tested these, but based on the well-documented commit notices
for Morphic-pre.1491 and Morphic-cmfcmf.1490, both should be moved/merged
to trunk.

The other inbox packages from Christoph Thiede appear to still be under
discussion on the list.

In the interest of reducing bit rot in the inbox, can anyone think of a
good reason that Morphic-pre.1491 and Morphic-cmfcmf.1490 should not go
into trunk now?

Thanks,
Dave


On Sun, Jun 30, 2019 at 10:08:05AM +0000, [hidden email] wrote:
> A new version of Morphic was added to project The Inbox:
> http://source.squeak.org/inbox/Morphic-cmfcmf.1490.mcz
>
> ==================== Summary ====================
>
> Name: Morphic-cmfcmf.1490
> Author: cmfcmf
> Time: 30 June 2019, 12:04:13.370226 pm
> UUID: ab2320d7-f8f4-df46-a786-ec28230e912c
> Ancestors: Morphic-pre.1489
>
> When wrapping a Morph into a flex shell, first add the flex shell to the Morph's owner (and thus to the Morph's world) before making the Morph a submorph of the flex shell.
> This avoids unnecessarily sending outOfWorld: to the Morph when adding it to the new flex shell (which doesn't have a world after creation) and intoWorld: when adding the flex shell to the Morph's owner right thereafter.
>
> =============== Diff against Morphic-pre.1489 ===============
>
> Item was changed:
> ----- Method: Morph>>addFlexShell (in category 'rotate scale and flex') -----
> addFlexShell
> "Wrap a rotating and scaling shell around this morph."
>
> | oldHalo myWorld flexMorph anIndex |
> -
> oldHalo:= self halo.
> myWorld := self world.
> self owner
> ifNil: [flexMorph := self newTransformationMorph asFlexOf: self]
> ifNotNil: [:myOwner |
> anIndex := myOwner submorphIndexOf: self.
> + "Avoid triggering outOfWorld: on self by first adding flexMorph to myOwner and only then making myself a submorph of flexMorph via asFlexOf:"
> + flexMorph := self newTransformationMorph.
> - flexMorph := self newTransformationMorph asFlexOf: self.
> myOwner addMorph: flexMorph asElementNumber: anIndex.
> + flexMorph asFlexOf: self.
> myWorld ifNotNil: [myWorld startSteppingSubmorphsOf: flexMorph]].
> self transferStateToRenderer: flexMorph.
> oldHalo ifNotNil: [oldHalo setTarget: flexMorph].
>
> ^ flexMorph!
>
>

Am 03.07.2019 23:37:18 schrieb Chris Muller <[hidden email]>:

On Tue, Jul 2, 2019 at 7:46 PM David T. Lewis <[hidden email]> wrote:

tl;dr Is there any reason NOT to move/merge Morphic-pre.1491 and
Morphic-cmfcmf.1490 from inbox to trunk?

Yes, this:

> I have not tested these, ...

I haven't had a chance either Dave, but given they're both less than a week old, I don't understand what your hurry is.

> but based on the well-documented commit notices
> for Morphic-pre.1491 and Morphic-cmfcmf.1490, both should be moved/merged
> to trunk. 

No, commit notices are not a substitute for testing and peer review.

Descending from Morphic-pre.1489, we have Morphic-pre.1490 that was moved
from inbox to treated, and we also have Morphic-cmfcmf.1490 still in inbox.
Parallel to that, we also have inbox Morphic-pre.1491, with inbox Morphic-ct.1492
descended from it, and Morphic-ct.1492 depending on inbox Tools-ct.851 (and
presumably also Compiler-ct.407 and Tests-ct.412.

Woops, I guess you missed Morphic-cmm.1489, which has been there since May.  Interesting...

 

I have not tested these, but based on the well-documented commit notices
for Morphic-pre.1491 and Morphic-cmfcmf.1490, both should be moved/merged
to trunk.

The other inbox packages from Christoph Thiede appear to still be under
discussion on the list.

In the interest of reducing bit rot in the inbox, can anyone think of a
good reason that Morphic-pre.1491 and Morphic-cmfcmf.1490 should not go
into trunk now?

They won't rot if you test them first.  Otherwise if there's a problem you could end up bringing cruft into the actual trunk, like your last trunk commit, which had to be rolled back.

 - Chris

 

Thanks,
Dave


On Sun, Jun 30, 2019 at 10:08:05AM +0000, [hidden email] wrote:
> A new version of Morphic was added to project The Inbox:
> http://source.squeak.org/inbox/Morphic-cmfcmf.1490.mcz
>
> ==================== Summary ====================
>
> Name: Morphic-cmfcmf.1490
> Author: cmfcmf
> Time: 30 June 2019, 12:04:13.370226 pm
> UUID: ab2320d7-f8f4-df46-a786-ec28230e912c
> Ancestors: Morphic-pre.1489
>
> When wrapping a Morph into a flex shell, first add the flex shell to the Morph's owner (and thus to the Morph's world) before making the Morph a submorph of the flex shell.
> This avoids unnecessarily sending outOfWorld: to the Morph when adding it to the new flex shell (which doesn't have a world after creation) and intoWorld: when adding the flex shell to the Morph's owner right thereafter.
>
> =============== Diff against Morphic-pre.1489 ===============
>
> Item was changed:
>   ----- Method: Morph>>addFlexShell (in category 'rotate scale and flex') -----
>   addFlexShell
>       "Wrap a rotating and scaling shell around this morph."
>   
>       | oldHalo myWorld flexMorph anIndex |
> -
>       oldHalo:= self halo.
>       myWorld := self world.
>       self owner
>               ifNil: [flexMorph := self newTransformationMorph asFlexOf: self]
>               ifNotNil: [:myOwner |
>                       anIndex := myOwner submorphIndexOf: self.
> +                     "Avoid triggering outOfWorld: on self by first adding flexMorph to myOwner and only then making myself a submorph of flexMorph via asFlexOf:"
> +                     flexMorph := self newTransformationMorph.
> -                     flexMorph := self newTransformationMorph asFlexOf: self.
>                       myOwner addMorph: flexMorph asElementNumber: anIndex.
> +                     flexMorph asFlexOf: self.
>                       myWorld ifNotNil: [myWorld startSteppingSubmorphsOf: flexMorph]].
>       self transferStateToRenderer: flexMorph.
>       oldHalo ifNotNil: [oldHalo setTarget: flexMorph].
>   
>       ^ flexMorph!
>
>